This proposal will examine the unanticipated hypothesis that the ?6?4 integrin (referred to as `?4') contributes to the initiation and metastasis of breast tumors by a non-cell autonomous mechanism. This hypothesis derives from the observations that the expression of ?4 is low or absent in breast cancer stem cells (CSCs) but that it contributes to the formation and metastasis of breast cancers. A major goal of this proposal is to reconcile these discrepant observations and understand how ?4 contributes to tumor formation and metastasis. Interestingly, triple-negative breast cancers (TNBCs) contain a distinct sub-population of cells characterized by high expression of the ?4 integrin (?4high) that is distinct from CSCs (?4-/low/CD24low/CD44high), exhibits basal epithelial differentiation and lacks stem cell properties. These observations indicate that repression of ?4 is necessary for CSC function. In fact, we observed unexpectedly that ?4 expression in CSCs promotes their differentiation and inhibits their self-renewal. The first specific aim will investigate the mechanism that underlies this phenomenon and it will focus on the exciting possibility that it involves the ability of ?4 to modulate focal adhesions and cytoskeletal tension resulting in diminished stem cell properties. This aim will also test the possibility that alterations in cytoskeletal tension have a causal role in promoting the differentiation of CSCs. The second specific aim will evaluate the hypothesis that the ?4high sub-population exerts non-cell autonomous regulation of CSCs. In other terms, a key function of the ?4high subpopulation is to expand the CSC population. More specifically, it is proposed that this distinct sub-population engages in paracrine signaling with CSCs mediated by parathyroid hormone-related protein (PTHrP) and that this signaling contributes to the expansion of CSCs. This hypothesis is based on the observations that ?4 regulates PTHrP expression and that the ?4high sub-population and CSCs are in proximity in breast tumors. Also, our preliminary data indicate that the ?4high sub-population helps to maintain self-renewal and survival. These observations provide an explanation for how the ?4 integrin contributes to tumor initiation and metastasis without being expressed in CSCs. The final specific aim will study the relationship between CSCs and ?4high cells in vivo and evaluate their relative contribution to tumor initiation and metastasis. The first set of experiments will examine whether both CSCs and ?4high cells are required for efficient tumor formation and metastasis using a transgenic model of TNBC. The second sub-aim probe more deeply into the relationship between CSCs and ?4high cells by evaluating the extent to which CSCs differentiate in vivo and assessing their association with ?4high cells in niches. These goals will be facilitated by the use of reporter constructs to tag these distinct populations. This approach will allow us to address several key issues including the frequency that CSCs (GFP+) differentiate in vivo, the association of ?4high and CSCs in `niches' and the relative frequency of CSCs and ?4high cells in metastases.